Math 403 Chapter 14: Ideals and Quotient (Factor) Rings 1. Introduction: In group theory we introduced the concept of a normal subgroup and we showed that if N/G then we can create the quotient (factor) group G=N. This idea has an analogy in the theory of rings. 2. Ideals: (a) Definition: A subring A ≤ R is called an ideal of R if 8r 2 R and 8a 2 A we have ar; ra 2 A. Definition: A is a proper ideal if it is an ideal which is not the entire ring. Example: For any ring R both f0g and R are ideals of R. Example: nZ is an ideal of Z. Example: The set of polynomials with real coefficients and constant term 0 is an ideal of R[x]. (b) Theorem (Ideal Test): If A ⊆ R with A 6= ; then A is an ideal of F if: i. 8a; b 2 A we have a − b 2 A. Note that a − b means a + (−b). ii. 8a 2 A and 8r 2 R we have ar; ra 2 A. Proof: This is a straightforward mash-up of the subring test and the definition of an ideal. QED (c) Definition: If R is a commutative ring with unity and a 2 R then the principal ideal generated by a is the set: hai = fra j r 2 Rg Note: The fact that its an ideal follows from the definition. Warning: We use the notation hgi in group theory but the definitions are different! Example: In R[x] the ideal hxi consists of all polynomials with constant term 0. Definition: We can expand the above for a1; :::; an 2 R commutative with unity to have the ideal generated by all of the ai: ha1; :::; ani = fr1a1 + ::: + rnan j r1; :::; rn 2 Rg Example: In Z[x] the ideal hx; 2i consists of all polynomials with even constant term. This is because an element in this ideal has the form p(x)(x)+q(x)(2) for p(x); q(x) 2 Z[x]. 3. Quotient (Factor) Rings: (a) Definition: Let R be a ring and A be a subring of A. Then the set of cosets (defined the same way as for groups with addition): R=A = fr + A j r 2 Rg is a ring under the operations (r + A) + (s + A) = (r + s) + A and (r + A)(s + A) = rs + A iff A is an ideal of R. Proof: Suppose A is an ideal of R. Since R is an Abelian group under addition we know A is a normal subgroup and so the set of cosets forms a group under addition. Next we need to show that our multiplication is well-defined. Suppose we have s + A = s0 + A and t + A = t0 + A, then s0 + 0 2 s + A so s0 = s + a and likewise t0 = t + b for a; b 2 A. Then observe that: s0t0 + A = (s + a)(t + b) + A = st + at + sb + ab + A = st + A The final equality holds because at; sb 2 A because A is an ideal and ab 2 A because A is a subring of R. Showing that multiplication is associative and distributes over addition follows immediately. Note that if A is not an ideal of R then choose a 2 A and r 2 R with (WLOG) ar 62 A. Then (a + A)(r + A) = ar + A but (a + A)(r + A) = (0 + A)(r + A) = 0 + A which contradicts the fact that ar 62 A. QED Example: The quotient ring Z=4Z consists of the elements f0+4Z; 1+4Z; 2+4Z; 3+4Zg with obvious operations, for example (2 + 4Z) + (3 + 4Z) = 5 + 4Z = 1 + 4Z and (2 + 4Z)(3 + 4Z) = 6 + 4Z = 2 + 4Z. 2 Example: Consider the quotient ring Z[x]= x − 2 . What do the distinct cosets look like? Well we have x2−2+ x2 − 2 = 0+ x2 − 2 so we can think of this as x2+ x2 − 2 = 2 + x2 − 2 . This allows drastic simplification of the cosets, for example: 7x6 + x5 − 3x2 + 4x − 1 + x2 − 2 = 7(x2)3 + (x2)2x − 3(x2) + 4x − 1 + x2 − 2 = 7(2)3 + (2)2x − 3(2) + 4x − 1 + x2 − 2 = 8x − 49 + x2 − 2 2 2 and similarly every p(x) + x − 2 is equivalent to ax + b + x − 2 with a; b 2 Z. Could we reduce further? In other words could two of these be the same? Well suppose ax+b+ x2 − 2 = cx+d+ x2 − 2 . Then we have (a−c)x+(b−d)+hx − 2i = 0+ x2 − 2 and so (a − c)x + (b − d) 2 x2 − 2 . 2 2 However elements in x − 2 have the form q(x)(x − 2) for q(x) 2 Z[x] and therefore have degree at least 2 except for the zero polynomial. Since (a − c)x + (b − d) has degree at most 1 it must be the zero polynomial and so a = c and b = d. Thus these elements are all distinct. How does multiplication work in this ring? In general: ax + b + x2 − 2 cx + d + x2 − 2 = acx2 + (ad + bc)x + bd + x2 − 2 = ac(2) + (ad + bc)x + bd + x2 − 2 = (ad + bc)x + (bd + 2ac) + x2 − 2 Example: Consider the quotient ring Z[i]= h3 + ii. What do the distinct cosets look like? Well we have 3+i+h3 + ii = 0+h3 + ii so we can think of this as i+h3 + ii = −3+h3 + ii. However since i2 = −1 we can square both sides to get −1 + h3 + ii = 9 + h3 + ii and so 10 + h3 + ii = 0 + h3 + ii. Since every coset has the form a + bi + h3 + ii such a coset can be rewritten by replacing i with −3 and 10 with 0, therefore every coset has the form c + h3 + ii for c 2 Z10. Could two of these coset be identical? Suppose a + h3 + ii = b + h3 + ii so that a − b 2 h3 + ii and so a − b = (c + di)(3 + i) = (3c − d) + (c + 3d)i for some c; d 2 Z. But then 3c − d = a − b and c + 3d = 0. Solving these yields a − b = −10d but since a; b 2 Z10 we have d = 0 and a = b. Thus they are unique. In fact this is essentially the ring Z10 written differently. 4. Maximal and Prime Ideals: (a) Definition: A proper ideal A of a commutative ring R is a maximal ideal of R if whenever B is another ideal with A ⊆ B ⊆ R then B = A or B = R. Basically this means that an ideal which is larger must be the entire ring. Typically proving that an ideal is maximal involves taking another ideal B with A ( B and showing B = R. Typically to show B = R we show 1 2 B because then r = r(1) 2 B for any r 2 R. Example: The ideal 6Z is not maximal in Z because 6Z( 2Z(Z. Example: The ideal 7Z is maximal in Z. To see this suppose 7Z( B ⊆ R, then there is some b 2 B with b 62 7Z and so gcd (7; b) = 1 and so there exist x; y 2 Z with 7x + by = 1. Then since b 2 B and 7 2 7Z( B we have 1 2 B and then r = r(1) 2 B for all r and so R = B. Example: The ideal hxi is not maximal in Z[x] since hxi ( hx; 2i (Z[x]. (b) Definition: A proper ideal A of a commutative ring R is a prime ideal of R if for all a; b 2 R if ab 2 A then a 2 A or b 2 A. Example: The ideal 6Z is not prime in Z because (2)(3) 2 6Z but 2 62 6Z and 3 62 6Z. Example: The ideal 7Z is prime in Z. To see this suppose ab 2 7Z. Then 7 j ab and so 7 j a or 7 j b and so a 2 7Z or b 2 7Z. Example: The ideal hxi is prime in Z[x]. To see this suppose p(x)q(x) 2 hxi. The ideal hxi consists of all polynomials with constant term zero and hence one of p(x) or q(x) must have constant term 0 since the constant term of p(x)q(x) is the product of the constant terms of p(x) and of q(x). Thus either p(x) or q(x) is in hxi. (c) Theorem: Let R be a commutative ring with unity and let A be an ideal. Then R=A is an integral domain iff A is a prime ideal. Proof: =): Suppose R=A is an integral domain and suppose ab 2 A. Then (a + A)(b + A) = ab + A = 0 + A so either a + A = 0 + A or b + A = 0 + A and so either a 2 A or b 2 A. (=: Suppose A is a prime ideal and suppose (a+A)(b+A) = 0+A. Then ab+A = 0+A and so ab 2 A and so either a 2 A or b 2 A and so either a + A = 0 + A or b + A = 0 + A. QED (d) Theorem: Let R be a commutative ring with unity and let A be an ideal. Then R=A is a field iff A is maximal. Proof: =): Suppose R=A is a field and A ( B ⊆ R. Let b 2 B with b 62 A. Then b + A 6= 0 + A and so b + A is a unit in R=A and so there is some c + A with (b + A)(c + A) = 1 + A.